5 ms·

Industrialization of Biotech vs. the Biologization of Industry

42 points by sebg 4y ago

jackcosgrove 4y ago
We kind of had "organic manufacturing" before the industrial era, in the form of skilled craftsmen. Two of the main problems with this approach are interoperability and reliability. Reliability seems like a solvable problem for biology, e.g. genetic propagation. Interoperability seems like a much tougher nut to crack.
philipkglass 4y ago
"The biologization of industry" seems to be complementary to the industrialization of biotech rather than an alternative to it. The pitch for biologization is: "Living in the industrialized world, we have giant retail stores that are packed with chemicals and detergents, medicines, and clothing and materials that have been shipped from all around the world in a complex supply chain. What if we didn’t need that? What if we could grow what we wanted locally? What if our supply chain was just biology?" That's only going to work for some materials. Bacteria are versatile but they can't produce materials like carbon fiber or copper alloys. They can help to produce inputs to making those materials, like acrylonitrile [1] or copper ore concentrates [2], but the fiber and alloys are going to be made in factories that still consume industrial quantities of material and need industrial scale supplies. [1] https://www.solvay.com/en/press-release/solvay-collaborate-trillium-bio-based-acrylonitrile-carbon-fiber-applications https://www.solvay.com/en/press-release/solvay-collaborate-t... [2] https://www.copper.org/publications/newsletters/innovations/2004/05/producing_copper_natures_way_bioleaching.html https://www.copper.org/publications/newsletters/innovations/...
- LarsDu88 4y ago
  We use copper to conduct electricity and carbon fiber to form rigid structures. Your body has parts that do both of these without using copper or carbon fiber, so your argument is not particularly compelling. Furthermore, who said we are limited to using bacteria?
throwawaymaths 4y ago
The biologization of industry is more or less correct, but we're not going to get antibody boxes (using current technology - yes there are some ideas kicking around and I'm thinking of someday doing a startup on it but initial capital expenses are high for industrial biotech and largely incompatible IMO with sv style vc); antibodies are not solely defined by their DNA sequencing and correctly manufacturing them is not simple.
trap_goes_hot 4y ago
Biology is not deterministic in the same way as traditional manufacturing is. Given the same input, repeating a biological process does not always produce identical results. When you body makes mistakes, it has other systems and pathways to address them. This isn't anything new, and is already well understood by domain experts. When it comes to making proteins using biological processes, product heterogeneity, especially, is a major concern that hasn't been fully eliminated - (aggregates, fragmentation, post-translation modifications, protein isomers, etc, etc).
- ackbar03 4y ago
  Is there any frontline research that addresses this you could refer me to? From my very limited knowledge, this is currently what is preventing biosimilar drugs from reducing the cost of expensive biologic drugs (e.g. Insulin and other enzyme based orphan drugs) unlike generics for regular drugs
  - mattkrause 4y ago
    Some keywords you might want to try include: - Replication error: copying DNA, transcribing it into RNA, and translating that into proteins uses processes that make mistakes. There are all sorts of interesting mechanisms for catching these, both in vivo and during sequencing. Try here for error rates: https://pubmed.ncbi.nlm.nih.gov/1805967/ https://pubmed.ncbi.nlm.nih.gov/1805967/ - Spontaneous mutation: A consequence of replication errors is that unwanted genomic changes sometimes arise in engineered organisms: you carefully make bacteria—-or a mouse—-with one mutation, but it develops another. Say “spontaneous mutation” around a large enough group of grad students and one will cry because it can derail experiments in hard-to-control ways https://pubmed.ncbi.nlm.nih.gov/ https://pubmed.ncbi.nlm.nih.gov/ Everything in biology is full of feedback loops and compensatory mechanisms. Knocking out one gene or brain region may have a direct effect on some output, but may also affect a bunch of other genes/regions (in many ways), which in turn…
  - trap_goes_hot 4y ago
    I am not aware of any research that is addressing heterogeneity issues in a general way. (Caveat: Not a formally trained scientist, I work in manufacturing) One of the products we manufacture is a protein that is expressed by an organism that we engineered (recombinant dna). The way we (and everyone else) achieve homogeneity is by purifying the protein. We develop this purification strategy at the research stage, after we characterize the protein (solubility, stability, size, hydrophobicity, etc). As 'research' turns into 'product', we iterate over this to improve yield, minimize processing times, lower raw material costs, etc and come up with an effective manufacturing process that we can make money on. I liked the article, but in my opinion, the 'personal biomaker' idea is something we are not close to implementing given the state of tech we currently have - at-least for proteins.
LarsDu88 4y ago
This essay uses too many analogies with the manufacturing of consumer electronic goods. I spent years using recombinant DNA to modify animals (nematodes and fruitflies), and I think a lot of folks don't appreciate the level of QC needed to manufacture and grow stuff at any level of scale. Your fingertips and the air itself is teeming with microorganisms that themselves are crawling with RNAses. My wife works in semiconductor manufacturing and the difference cannot be more stark. I went to a science conference with her and damn near every chart was filled with straight lines that differed no more than nanometers. I'm now a software engineer (still working in biotech), and I can tell you there is a lazy level of comfort in being able to type code and get the same result every time
- kwere 4y ago
  QC?
  - jjk166 4y ago
    Quality control
- washadjeffmad 4y ago
  Agreed that there are no real quality or VLSI engineering parallels with biologics. Spontaneous reversion of mutations, natural mutations, lateral gene transfer, and more are always barriers. Lab work is fun because it always leads to more questions, but it makes the precision of computing only feel like theory crafting. It's hard to scale in biotech, which does reduce some burden on efficiency. Processes that work at the bench and pilot scale may not work in larger reactors or processes, or only inconsistently. Custom designs for each strain or species would be ideal, but that isn't generally taken very far because it's easier to focus on costs and yields. Whatever can be done most cheaply or the fastest with what already exists is often what wins out. It's funny, I worked on media development and reactor design for nematode cultures over a decade ago, which I moved to from programming. After frustration with diminishing grants and research budgets, I transitioned to IT, and I'm considering a move to IS to support some projects I care about. The way HN stories often end with "and then I became a software developer and never looked back" makes me feel like I've had my career in reverse :)
Protostome 4y ago
I don't see biology becoming a consumer business anytime soon (i.e. printing molecules automatically in your garage). As some remarked, biology is inherently chaotic, a slight change in reaction condition may yield totally different result.

42 points by sebg 4y ago

We kind of had "organic manufacturing" before the industrial era, in the form of skilled craftsmen. Two of the main problems with this approach are interoperability and reliability. Reliability seems like a solvable problem for biology, e.g. genetic propagation. Interoperability seems like a much tougher nut to crack.

philipkglass 4y ago

"The biologization of industry" seems to be complementary to the industrialization of biotech rather than an alternative to it. The pitch for biologization is: "Living in the industrialized world, we have giant retail stores that are packed with chemicals and detergents, medicines, and clothing and materials that have been shipped from all around the world in a complex supply chain. What if we didn’t need that? What if we could grow what we wanted locally? What if our supply chain was just biology?" That's only going to work for some materials. Bacteria are versatile but they can't produce materials like carbon fiber or copper alloys. They can help to produce inputs to making those materials, like acrylonitrile [1] or copper ore concentrates [2], but the fiber and alloys are going to be made in factories that still consume industrial quantities of material and need industrial scale supplies. [1] https://www.solvay.com/en/press-release/solvay-collaborate-trillium-bio-based-acrylonitrile-carbon-fiber-applications https://www.solvay.com/en/press-release/solvay-collaborate-t... [2] https://www.copper.org/publications/newsletters/innovations/2004/05/producing_copper_natures_way_bioleaching.html https://www.copper.org/publications/newsletters/innovations/...

LarsDu88 4y ago
We use copper to conduct electricity and carbon fiber to form rigid structures. Your body has parts that do both of these without using copper or carbon fiber, so your argument is not particularly compelling. Furthermore, who said we are limited to using bacteria?

throwawaymaths 4y ago

The biologization of industry is more or less correct, but we're not going to get antibody boxes (using current technology - yes there are some ideas kicking around and I'm thinking of someday doing a startup on it but initial capital expenses are high for industrial biotech and largely incompatible IMO with sv style vc); antibodies are not solely defined by their DNA sequencing and correctly manufacturing them is not simple.

trap_goes_hot 4y ago

Biology is not deterministic in the same way as traditional manufacturing is. Given the same input, repeating a biological process does not always produce identical results. When you body makes mistakes, it has other systems and pathways to address them. This isn't anything new, and is already well understood by domain experts. When it comes to making proteins using biological processes, product heterogeneity, especially, is a major concern that hasn't been fully eliminated - (aggregates, fragmentation, post-translation modifications, protein isomers, etc, etc).

ackbar03 4y ago
Is there any frontline research that addresses this you could refer me to? From my very limited knowledge, this is currently what is preventing biosimilar drugs from reducing the cost of expensive biologic drugs (e.g. Insulin and other enzyme based orphan drugs) unlike generics for regular drugs
- mattkrause 4y ago
  Some keywords you might want to try include: - Replication error: copying DNA, transcribing it into RNA, and translating that into proteins uses processes that make mistakes. There are all sorts of interesting mechanisms for catching these, both in vivo and during sequencing. Try here for error rates: https://pubmed.ncbi.nlm.nih.gov/1805967/ https://pubmed.ncbi.nlm.nih.gov/1805967/ - Spontaneous mutation: A consequence of replication errors is that unwanted genomic changes sometimes arise in engineered organisms: you carefully make bacteria—-or a mouse—-with one mutation, but it develops another. Say “spontaneous mutation” around a large enough group of grad students and one will cry because it can derail experiments in hard-to-control ways https://pubmed.ncbi.nlm.nih.gov/ https://pubmed.ncbi.nlm.nih.gov/ Everything in biology is full of feedback loops and compensatory mechanisms. Knocking out one gene or brain region may have a direct effect on some output, but may also affect a bunch of other genes/regions (in many ways), which in turn…
- trap_goes_hot 4y ago
  I am not aware of any research that is addressing heterogeneity issues in a general way. (Caveat: Not a formally trained scientist, I work in manufacturing) One of the products we manufacture is a protein that is expressed by an organism that we engineered (recombinant dna). The way we (and everyone else) achieve homogeneity is by purifying the protein. We develop this purification strategy at the research stage, after we characterize the protein (solubility, stability, size, hydrophobicity, etc). As 'research' turns into 'product', we iterate over this to improve yield, minimize processing times, lower raw material costs, etc and come up with an effective manufacturing process that we can make money on. I liked the article, but in my opinion, the 'personal biomaker' idea is something we are not close to implementing given the state of tech we currently have - at-least for proteins.

LarsDu88 4y ago

This essay uses too many analogies with the manufacturing of consumer electronic goods. I spent years using recombinant DNA to modify animals (nematodes and fruitflies), and I think a lot of folks don't appreciate the level of QC needed to manufacture and grow stuff at any level of scale. Your fingertips and the air itself is teeming with microorganisms that themselves are crawling with RNAses. My wife works in semiconductor manufacturing and the difference cannot be more stark. I went to a science conference with her and damn near every chart was filled with straight lines that differed no more than nanometers. I'm now a software engineer (still working in biotech), and I can tell you there is a lazy level of comfort in being able to type code and get the same result every time

kwere 4y ago
QC?
- jjk166 4y ago
  Quality control
washadjeffmad 4y ago
Agreed that there are no real quality or VLSI engineering parallels with biologics. Spontaneous reversion of mutations, natural mutations, lateral gene transfer, and more are always barriers. Lab work is fun because it always leads to more questions, but it makes the precision of computing only feel like theory crafting. It's hard to scale in biotech, which does reduce some burden on efficiency. Processes that work at the bench and pilot scale may not work in larger reactors or processes, or only inconsistently. Custom designs for each strain or species would be ideal, but that isn't generally taken very far because it's easier to focus on costs and yields. Whatever can be done most cheaply or the fastest with what already exists is often what wins out. It's funny, I worked on media development and reactor design for nematode cultures over a decade ago, which I moved to from programming. After frustration with diminishing grants and research budgets, I transitioned to IT, and I'm considering a move to IS to support some projects I care about. The way HN stories often end with "and then I became a software developer and never looked back" makes me feel like I've had my career in reverse :)

Protostome 4y ago

I don't see biology becoming a consumer business anytime soon (i.e. printing molecules automatically in your garage). As some remarked, biology is inherently chaotic, a slight change in reaction condition may yield totally different result.